You are on page 1of 9

Articles

Aggressive medical treatment with or without stenting in high-risk patients with intracranial artery stenosis (SAMMPRIS): the nal results of a randomised trial
Colin P Derdeyn*, Marc I Chimowitz*, Michael J Lynn, David Fiorella, Tanya N Turan, L Scott Janis, Jean Montgomery, Azhar Nizam, Bethany F Lane, Helmi L Lutsep, Stanley L Barnwell, Michael F Waters, Brian L Hoh, J Maurice Hourihane, Elad I Levy, Andrei V Alexandrov, Mark R Harrigan, David Chiu, Richard P Klucznik, Joni M Clark, Cameron G McDougall, Mark D Johnson, G Lee Pride Jr, John R Lynch, Osama O Zaidat, Zoran Rumboldt, Harry J Cloft, for the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis Trial Investigators

Summary
Background Early results of the Stenting and Aggressive Medical Management for Preventing Recurrent stroke in Intracranial Stenosis trial showed that, by 30 days, 33 (147%) of 224 patients in the stenting group and 13 (58%) of 227 patients in the medical group had died or had a stroke (percentages are product limit estimates), but provided insucient data to establish whether stenting oered any longer-term benet. Here we report the long-term outcome of patients in this trial. Methods We randomly assigned (1:1, stratied by centre with randomly permuted block sizes) 451 patients with recent transient ischaemic attack or stroke related to 7099% stenosis of a major intracranial artery to aggressive medical management (antiplatelet therapy, intensive management of vascular risk factors, and a lifestyle-modication programme) or aggressive medical management plus stenting with the Wingspan stent. The primary endpoint was any of the following: stroke or death within 30 days after enrolment, ischaemic stroke in the territory of the qualifying artery beyond 30 days of enrolment, or stroke or death within 30 days after a revascularisation procedure of the qualifying lesion during follow-up. Primary endpoint analysis of between-group dierences with log-rank test was by intention to treat. This study is registered with ClinicalTrials.gov, number NCT 00576693. Findings During a median follow-up of 324 months, 34 (15%) of 227 patients in the medical group and 52 (23%) of 224 patients in the stenting group had a primary endpoint event. The cumulative probability of the primary endpoints was smaller in the medical group versus the percutaneous transluminal angioplasty and stenting (PTAS) group (p=00252). Beyond 30 days, 21 (10%) of 210 patients in the medical group and 19 (10%) of 191 patients in the stenting group had a primary endpoint. The absolute dierences in the primary endpoint rates between the two groups were 71% at year 1 (95% CI 02 to 138%; p=00428), 65% at year 2 (05 to 135%; p=007) and 90% at year 3 (15 to 165%; p=00193). The occurrence of the following adverse events was higher in the PTAS group than in the medical group: any stroke (59 [26%] of 224 patients vs 42 [19%] of 227 patients; p=00468) and major haemorrhage (29 [13%] of 224 patients vs 10 [4%] of 227 patients; p=00009). Interpretation The early benet of aggressive medical management over stenting with the Wingspan stent for highrisk patients with intracranial stenosis persists over extended follow-up. Our ndings lend support to the use of aggressive medical management rather than PTAS with the Wingspan system in high-risk patients with atherosclerotic intracranial arterial stenosis. Funding National Institute of Neurological Disorders and Stroke (NINDS) and others.
Published Online October 26, 2013 http://dx.doi.org/10.1016/ S0140-6736(13)62038-3 See Online/Comment http://dx.doi.org/10.1016/ S0140-6736(13)62191-1 *Authors contributed equally Members listed in the appendix Mallinckrodt Institute of Radiology and the Departments of Neurology and Neurosurgery, Washington University School of Medicine, St Louis, MO, USA (Prof C P Derdeyn MD); Department of Neurosciences, Medical University of South Carolina, Charleston, SC, USA (Prof M I Chimowitz MBChB, T N Turan MD); Department of Biostatistics and Bioinformatics, Emory University Rollins School of Public, Health, Atlanta, GA, USA (M J Lynn MS, J Montgomery RN, B F Lane RN, A Nizam MS); Department of Neurosurgery, State University of New York, Stony Brook, NY, USA (Prof D Fiorella MD); National Institute of Neurological Disorders and Stroke, National Institute of Health, Bethesda, MD, USA (L S Janis PhD); Department of Neurology, Oregon Health and Science University, Portland, OR, USA (Prof H L Lutsep MD); Department of Neurological Surgery and the Dotter Interventional Institute, Oregon Health Sciences University, Portland, OR, USA (S L Barnwell MD); Departments of Neurology and Neuroscience (M F Waters MD) and Department of Neurosurgery (B L Hoh MD) University of Florida, Gainesville, FL, USA; Department of Neurosurgery, University of Bualo, NY, USA (Prof E I Levy MD); Dent Neurological Institute, Bualo,

Introduction
Intracranial atherosclerosis is a common cause of stroke and is associated with a high risk of recurrent stroke, especially in patients with a recent stroke or transient ischaemic attack and severe arterial stenosis.14 The Stenting and Aggressive Medical Management for Preventing Recurrent stroke in Intracranial Stenosis (SAMMPRIS) trial was designed to assess whether percutaneous transluminal angioplasty and stenting (PTAS) plus aggressive medical treatment is more eective than aggressive medical treatment alone in high-risk patients with this disease.5 Enrolment in

SAMMPRIS began on Nov 25, 2008, but was stopped for safety concerns on April 5, 2011, because the 30-day rate of stroke and death was higher in the PTAS group.6 When enrolment was stopped, fewer than half the 451 patients had been followed up for longer than 1 year.6 Since then, patients in both treatment groups have been followed up for 2 more years to establish whether the early benet in the medical group would persist over longer follow-up, or whether the medical group would have a high incidence of late strokes that would eliminate the early ecacy gap between the two groups. In this Article, we report the nal results of the SAMMPRIS trial.

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

NY, USA (J M Hourihane MD); Department of Neurology (Prof A V Alexandrov MD), and Department of Neurosurgery (M R Harrigan MD), University of Alabama, Birmingham, AL, USA; Department of Neurology (Prof D Chiu MD), and Department of Radiology (R P Klucznik MD), Houston Methodist Hospital, Houston, TX, USA; Department of Neurology (J M Clark MD), and Department of Neurosurgery (Prof C G McDougall MD), Barrow Neurological Institute, Phoenix, AZ, USA; Department of Neurology and Neurotherapeutics (M D Johnson MD), Departments of Radiology and Neurosurgery (Lee Pride Jr MD), University of Texas Southwestern Medical Center, Dallas, TX, USA; Department of Neurology, Medical College of Wisconsin, Milwaukee, WI, USA (J R Lynch MD); Departments of Neurology, Radiology, and Neurosurgery, Medical College of Wisconsin, Milwaukee, WI (Prof O O Zaidat MD); Department of Radiology, Medical University of South Carolina, Charleston, SC, USA (Prof Z Rumboldt MD); and Department of Radiology, Mayo Clinic, Rochester, MN, USA (Prof H J Cloft MD) Correspondence to: Prof Colin P Derdeyn, Washington University School of Medicine/ Barnes Jewish Hospital, 510 South Kingshighway Boulevard, St Louis, MO 63110, USA derdeync@wustl.edu

Methods
Patients and study design
The trial design and early results are available elsewhere.57 SAMMPRIS was a randomised, superiority, multi-centre, clinical trial funded by the National Institute of Neurological Disorders and Stroke (NINDS). The US Food and Drug Administration (FDA) issued an Investigational Device Exemption to do the study with the Wingspan stent system (Stryker Neurovascular, Fremont, CA, USA; formerly Boston Scientic Neurovascular),8 which had been approved under a Humanitarian Device Exemption in 2005 for patients with 5099% intracranial stenosis who are refractory to medical treatment.9 Eligible patients were 3080 years old and had non-disabling stroke or transient ischaemic attack within 30 days before enrolment that was attributed to 7099% atherosclerotic stenosis of a major intracranial artery veried by catheter angiography.10 Exclusion criteria included tandem extracranial or intracranial stenosis (7099%) proximal or distal to the target intracranial stenosis, intraluminal thrombus proximal to or at the target lesion, progressive neurological signs within 24 h before enrolment, any haemorrhagic infarct within 14 days before enrolment, non-atherosclerotic causes of intracranial stenosis, and the presence of a cardiac source of embolus. Other eligibility criteria are provided in the study protocol. All patients gave written informed consent to participate. Institutional review boards at all 50 participating sites in the USA approved the study protocol.

number generator with randomly permuted block sizes.5 The study participants, coordinators, and primary neurologists were not masked to treatment assignment but a second site neurologist was masked to assess patients with mild stroke endpoints. All potential neurological, cardiac, and haemorrhagic endpoints were adjudicated by independent panels of neurologists and cardiologists who were masked to treatment.

Treatments
Details of the implementation of aggressive medical management have been reported elsewhere.7 Briey, with the exception of the peri-procedural antithrombotic treatment in the PTAS group, medical management was identical in both groups throughout the entire follow-up period. Treatment consisted of aspirin (325 mg per day) for the duration of follow-up, clopidogrel (75 mg per day) for 90 days after enrolment, management of the primary risk factors (systolic blood pressure [SBP] and low-density lipoprotein cholesterol [LDL-C]), targeting SBP lower than 140 mm Hg (<130 mm Hg if diabetic) and LDL-C lower than 181 mmol/L, and management of secondary risk factors (diabetes, non-high-density lipoprotein cholesterol, smoking, weight, exercise) with the help of a lifestyle modication programme.11 We provided aspirin, clopidogrel, one drug from each major class of antihypertensive agents, and rosuvastatin. The lifestyle modication programme was delivered by a coach who contacted the patients by telephone every 2 weeks for the rst 3 months after enrolment and then monthly. The programme included baseline assessment of risk factors, formulating an action plan (with a focus on exercise, correct nutrition, weight management, and tobacco smoking cessation), and follow-up counselling sessions to help patients acquire the skills, motivation, and support needed to implement and adhere to their individualised action plan. Written health education materials were also provided to the study patients. PTAS was done by neurointerventionists who were credentialed to participate in the trial based on review of 20 consecutive intracranial angioplasty and stenting cases, at least three of which must have used the Wingspan system. Details on the credentialling process, monitoring of PTAS performance during the trial, and peri-procedure outcomes according to previous operator experience are available elsewhere.5,12 A 600 mg dose of clopidogrel was given to patients 624 h before PTAS if they had not been taking clopidogrel 75 mg daily for the previous 5 days. Intravenous heparin was used during the procedure to maintain an activated clotting time of 250300 s. Other details of the procedure and post-PTAS care are provided in the study protocol.

Randomisation and masking


We used a one-to-one ratio to randomly allocate patients to treatment, stratifying by centre. The randomisation sequence was produced at the Statistical Coordinating Center (Department of Biostatistics and Bioinformatics, Emory University, GA, USA), using a pseudo-random
451 patients randomised

227 patients assigned to medical treatment only 218 medical treatment only 9 underwent PTAS during follow-up*

224 patients assigned to medical treatment plus PTAS 208 stent placed 5 angioplasty only 7 procedure aborted before lesion accessed 4 procedure not done

11 patients lost to follow-up 13 patients withdrew consent

7 patients lost to follow-up 3 patients withdrew consent

227 patients analysed

224 patients analysed

Duration of follow-up and outcomes


The original plan was to follow patients up for a maximum of 3 years and a minimum of 1 year, but when enrolment in the trial was stopped a decision was made

Figure 1: Trial prole PTAS=percutaneous transluminal angioplasty and stenting. *Months after enrolment when PTAS was done: 01, 09, 22, 26, 32, 38, 65, 82, 141. See appendix for length of follow-up for these patients.

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

by the Executive Committee (MIC, CPD, MJL, DF, TNT, LSJ, JM, and BFL), NINDS, and the data and safety monitoring board to extend the follow-up of all patients to a common termination date 2 years after the last patient was enrolled. We assessed patients at study entry, 4 days, 30 days, and then every 4 months. All follow-up visits were in person unless the patient could not return for the visit, in which case telephone follow-up was done. Every in-person visit included blood pressure measurement, an examination by the study neurologist, and management of the patients risk factors. The schedule for doing laboratory tests, the National Institute of Health Stroke (NIH) stroke scale, the modied Rankin scale, and the Barthel index are provided in the study protocol. Telephone follow-up was done by the site coordinator and included assessment of the patients functional abilities and mobility using the Barthel index and modied Rankin scale, and establishing whether the patient had any adverse events. If so, we obtained the patients medical records for those events for review. Patients with a suspected stroke underwent brain MRI or CT and were assessed by the primary site neurologist. Patients who had events that were potentially dicult to classify (a transient ischaemic attack lasting >1 h or mild ischaemic stroke [an increase in the patients NIH stroke scale of <4 from study entry]) were assessed by a second site neurologist who was masked to treatment. Both neurologists assessments were sent for central adjudication. The primary endpoint was any of the following: any stroke or death within 30 days after enrolment, ischaemic stroke in the territory of the qualifying artery beyond 30 days of enrolment, or any stroke or death within 30 days after a revascularisation procedure of the qualifying lesion during follow-up (ie, stenting of a patient in the medical group or angioplasty for symptomatic restenosis in a patient in the PTAS group). We dened ischaemic stroke as a new focal neurological decit of sudden onset, lasting for at least 24 h, that was not associated with a haemorrhage on brain CT or MRI. We categorised ischaemic strokes as either in or out of the territory of the qualifying artery. We dened symptomatic brain haemorrhage as parenchymal, subarachnoid, or intraventricular haemorrhage detected by CT or MRI that was associated with new neurological signs or symptoms lasting 24 h or more or a seizure.

size, probability of type I error of 005, power of 080, 5% crossover rate from the medical to the PTAS group, and 2% loss to follow-up, the intended sample size based on a two-sided log-rank test was 382 patients per group. The log-rank test was used to compare the treatment groups for the primary and secondary endpoints. Cumulative probabilities based on estimates of the primary endpoint from the Kaplan-Meier curves at specic timepoints (years 1, 2, and 3) were compared between the treatment groups with a Z test. We
Medical group (N=227) Age in years Men Ethnic origin (self-reported) Black White Other History of hypertension History of lipid disorder Smoking Never Previously Currently Diabetes Systolic blood pressure (mm Hg) Low-density lipoprotein cholesterol (mmol/L) High-density lipoprotein cholesterol (mmol/L) Non-high-density lipoprotein cholesterol (mmol/L) Glycosylated haemoglobin in patients with diabetes Body-mass index (kg/m) History of coronary artery disease History of stroke (not qualifying event) Qualifying event Stroke Transient ischaemic attack On antithrombotic therapy at qualifying event Days from qualifying event to randomisation (median [IQR]) Symptomatic qualifying artery Internal carotid artery Middle cerebral artery Vertebral artery Basilar artery Percent stenosis of symptomatic qualifying artery Categories of percent stenosis of symptomatic qualifying artery 7079% 8089% 9099% 102 (45%) 97 (43%) 28 (12%) 49 (22%) 105 (46%) 22 (10%) 51 (22%) 81 (7) 152 (67%) 75 (33%) 140 (62%) 7 (419) 78 (34%) 80 (35%) 69 (30%) 103 (45%) 1468 (218) 253 (095) 100 (026) 302 (104) 82 (23) 307 (63) 59 (26%) 58 (26%) 49 (22%) 162 (71%) 16 (7%) 203 (89%) 202 (89%) 595 (118) 145 (64%)

For the study protocol see http://sammpris.org/ protocolsammpris.html

PTAS group (N=224) 610 (107) 127 (57%) 55 (25%) 160 (71%) 9 (4%) 200 (89%) 195 (87%) 90 (40%)* 79 (35%)* 54 (24%)* 105 (47%) 1439 (206) 249 (099) 098 (027) 301 (114) 78 (21) 303 (62) 47 (21%) 60 (27%) 142 (63%) 82 (37%) 144 (64%) 7 (416) 45 (20%) 92 (41%) 38 (17%) 49 (22%) 80 (7) 107/223 (48%) 92/223 (41%) 24/223 (11%)

Statistical methods
The study was originally designed to detect a 35% relative risk reduction in the primary endpoint with PTAS compared with a projected rate in the medical group of 247% at 2 years.5,6 This rate was based on the outcome of patients with recent symptoms and 7099% intracranial arterial stenosis who participated in the WASID trial13 and incorporated a proposed 15% relative risk reduction with aggressive medical management. With this eect

Data are mean (SD) or n (%), unless otherwise stated. PTAS=percutaneous transluminal angioplasty and stenting. *Data for one patient missing, so denominator is 223. Diabetic at baseline if history of diabetes or if haemoglobin A1c 65%. For this analysis, baseline haemoglobin A1c values for any patient with a history of diabetes, new diagnosis of diabetes during follow-up, or haemoglobin A1c 65% at any time in the study were included. According to reading of angiogram by site interventionist.

Table 1: Baseline characteristics

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

A
100 90 80 Proportion of patients (%) 70 60 50 40 30 20 10 0 Medical management PTAS

B
100 90 80 Proportion of patients (%) 70 60 50 40 30 20 10 0 0 4 8 12 16 20 Months of follow-up 24 28 32 36

Figure 2: Patients achieving target systolic blood pressure (A) and low-density lipoprotein cholesterol (B) throughout the trial (A) Target systolic blood pressure was <140 mm Hg (<130 mm Hg if diabetic). (B) Target low-density lipoprotein was <181 mmol/L. PTAS=percutaneous transluminal angioplasty and stenting.

investigated longitudinal dierences in risk factors between the treatment groups using generalised estimating equation regression models with an exchangeable correlation structure. Subgroup analyses for prespecied baseline factors were done by tting a Cox proportional hazards regression model that included treatment, a treatment by time interaction (because the proportional hazards assumption for the treatment was not met), the factor, and the treatment by factor interaction (the p value for which is reported in this paper). All analyses were done by intention to treat. All reported p values are two-sided without adjustment for multiple testing. We did three prespecied sensitivity analyses to generate data for patients withdrawing or becoming lost to follow-up. First, all patients lost to follow-up or withdrawn were considered to have had a primary endpoint at last contact. Second, only patients in the PTAS group were considered to have had a primary endpoint at last contact. Third, only patients in the medical group were considered to have had a primary endpoint at last contact. Additionally, we used a simulation approach (appendix) in which a hazard rate per month for the primary endpoint was estimated for each treatment group from available data for all patients with patients lost to follow-up or withdrawn censored at their last study visit. For each patient lost to follow-up or withdrawn, we randomly determined whether a primary endpoint occurred for each study month according to the hazard rate. If a simulated primary endpoint did not occur in any month, the patient was censored at the longest possible follow-up time. The simulated data for patients lost to follow-up or withdrawn were combined with the actual data for the remaining patients and the p value for the log-rank test comparing the treatment groups and the probabilities of a primary endpoint by 2 years were computed. The process was repeated 10 000 times. We used SAS (version 9.3) for all statistical analyses.

Role of the funding source


Low-density lipoprotein cholesterol (mmol/L) Medical group n Baseline 30 days 4 months 1 year 2 years 3 years 226 215 193 138 95 39 Mean (SD) 253 (095) 185 (069) 184 (067) 175 (071) 161 (058) 174 (066) PTAS group n 220 210 186 138 100 45 Mean (SD) 249 (099) 201 (073) 196 (103) 179 (076) 174 (080) 174 (079) Systolic blood pressure (mm Hg) Medical group n 227 216 198 172 149 88 Mean (SD) 1468 (218) 1408 (217) 1342 (170) 1321 (161) 1291 (141) 1287 (163) PTAS group n 220 213 186 172 153 86 Mean (SD) 1439 (206) 1388 (200) 1331 (158) 1295 (174) 1300 (154) 1267 (142)

The Executive Committee designed the study and oversaw trial activities. A representative of the study sponsor was on the Executive Committee of the trial (LSJ) and participated in oversight of the trial. Industry partners supplied the stents and rosuvastatin for study patients and supplemental funding for third-party monitoring and auditing of sites, but did not participate in the design, conduct, data analysis, or reporting of the trial. CPD, MIC, and MJL had full access to the study data and had nal responsibility to submit for publication.

For low-density lipoprotein cholesterol at years 1, 2, and 3, we used the value that was closest to the designated timepoint within a window of 4 months before or after the timepoint. For systolic blood pressure, the value that was closest to the designated timepoint within a window of 2 months before or after the timepoint was used. PTAS=percutaneous transluminal angioplasty and stenting.

Results
Between Nov 25, 2008, and March 31, 2011, we randomly assigned 451 patients to treatment: 227 to the medical group and 224 to the PTAS group (gure 1).Baseline characteristics were much the same between the two

Table 2: Primary risk factors according to treatment and visit

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

Medical group (N=227) Patients with Probability at 1 year event Primary endpoint Any stroke or death Any death Any stroke Disabling or fatal stroke Myocardial infarction Major non-stroke haemorrhage Any major haemorrhage 34 (15%) 51 (23%) 13 (6%) 42 (19%) 18 (8%) 9 (4%) 9 (4%) 10 (4%) 126% (89%178%) 175% (131%232%) 40% (20%78%) 149% (108%203%) 67% (40%111%) 34% (16%71%) 19% (07%51%) 24% (10%57%) Probability at 2 years 141% (101%194%) 198% (151%256%) 45% (24%85%) 172% (129%229%) 78% (48%124%) 34% (16%71%) 30% (14%67%) 35% (17%72%)

PTAS group (N=224) Patients with event 52 (23%) 64 (29%) 13 (6%) 59 (26%) 21 (9%) 5 (2%) 16 (7%) 29 (13%) Probability at 1 year 197% (150%255%) 228% (178%289%) 40% (21%80%) 219% (170%279%) 91% (59%138%) 21% (08%55%) 44% (23%83%) 98% (65%146%) Probability at 2 years 206% (159%265%) 241% (191%303%) 46% (24%87%) 233% (183%294%) 101% (67%151%) 27% (11%64%) 79% (48%127%) 131% (91%185%)

p value*

00252 013 090 00468 051 034 011 00009

Data are n (%) or probability (95% CI). PTAS=percutaneous transluminal angioplasty and stenting. *The p value is for comparison of time-to-event curves for the two treatment groups for each of the specied adverse events, calculated with the log-rank test. Causes of death in the medical group=brain haemorrhage (n=1), ischaemic stroke (n=1), myocardial infarction (n=2; one had an ischaemic stroke in the territory 17 days before but the immediate cause of death was myocardial infarction), cardiac arrest after coronary artery bypass surgery (n=1), sudden cardiac death (n=1), and non-vascular death (n=7); causes of death in the PTAS group=brain haemorrhage (n=4), ischaemic stroke (n=2), sudden cardiac death (n=2), and non-vascular death (n=5). Disability was based on the post-stroke assessment closest to 90 days after a patients stroke (a disabling stroke was dened as any of the following: modied Rankin of 4, Barthel index of 80, composite National Institutes of Health (NIH) stroke scale of 7), NIH stroke scale of 3 for motor examination of arm or leg, NIH stroke scale of 2 for best language, NIH stroke scale of 3 for visual category ie, cortical blindness; a stroke was considered fatal for this analysis if death occurred within 30 days or if stroke was the adjudicated cause of death. Any subdural or epidural haemorrhage or a systemic haemorrhage requiring any of the following: admission to hospital, blood transfusion, or surgery. Any intracranial haemorrhage or a systemic haemorrhage requiring any of the following: admission to hospital, blood transfusion, or surgery; one patient in the PTAS group had both a symptomatic brain haemorrhage and a major non-stroke haemorrhage.

Table 3: Comparison of treatment groups for the primary endpoint, secondary endpoints, and other major adverse events

groups (table 1). The median duration of follow-up in all patients was 324 months (IQR 242405; range 0 months to 526 months). A larger proportion of patients in the medical group were lost to follow-up or withdrew consent than in the PTAS group (24 [11%] of 227 vs 10 [5%] of 224; p=00193). Most withdrawals in both groups occurred after 1 year of follow-up (appendix). The percentages of patients who achieved SBP and LDL-C targets and the mean SBP and LDL-C values at various stages of follow-up in both groups are shown in gure 2 and table 2. Data for the secondary risk factors are shown in the appendix. In these analyses, we detected no statistically signicant between-group difference over time for any risk factors when analysed in terms of either the actual values or the percentage of patients in target range. Throughout the trial, primary endpoint events occurred more often in the PTAS group than in the medical group (table 3, gure 3). The occurrence of primary endpoints in the medical group versus PTAS group was 58% versus 147% at day 30 (p=00016), 126% versus 197% at year 1 (p=00428), 141% versus 206% at year 2 (p=007), and 149% versus 239% at year 3 (p=00193). Beyond 30 days of enrolment, 10% of patients in each group had a primary endpoint event (21 of 210 patients in the medical group and 19 of 191 patients in the PTAS group; appendix). Beyond 1 year of enrolment, six (3%) of 185 patients in the medical group and eight (5%) of 175 patients in the PTAS group had a primary endpoint event. The results of all sensitivity analyses are provided in the appendix. The worst case scenario for the medical group in which only patients in the medical group who withdrew or were lost to follow-up were assigned primary endpoints at last contact still resulted in the

10 Cumulative probability of the primary endpoint 09 08 07 06 05 04 03 02 01 0 0 Number at risk Medical group 227 PTAS group 224 6 199 184 12 185 175 18 24 30 36 Months since randomisation 172 170 132 128 92 91 42 47 50 p=00252

Medical group PTAS group

48 12 13

54

180 173

Figure 3: Cumulative probability of a primary endpoint by treatment PTAS=percutaneous transluminal angioplasty and stenting.

probability of a primary endpoint being numerically higher in the PTAS group, although not statistically signicantly higher. In the simulation analysis, none of the 10 000 repetitions resulted in the probability of a primary endpoint being less common in the PTAS group than in the medical group. Table 3 shows the secondary endpoints and other major adverse events in each group (additional details on all events in appendix). The rates of any stroke (p=00468) and any major haemorrhage (p=00009) were statistically signicantly lower in the medical group than they were in the PTAS group. Results of subgroup analyses for pre-specied baseline factors with rates of the primary endpoint at

See Online for appendix

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

Medical group n Age (median) <60 years 60 years Gender Male Female Symptomatic artery ICA MCA Vertebral Basilar Qualifying event TIA Stroke 7 >7 No Yes On antithrombotic at time of qualifying event No Yes Old infarct in territory and stroke as qualifying event while on antithrombotic treatment No Yes 186 36 10% 35% 87 140 12% 16% 75 152 115 112 147 75 7% 18% 13% 15% 9% 23% 49 105 22 51 23% 13% 10% 10% 145 82 11% 20% 119 108 12% 17% Probability by 2 years*

PTAS group n Probability by 2 years*

p value

075 107 117 127 97 45 92 38 49 82 142 115 109 146 69 80 144 16% 25% 040 18% 24% 063 29% 14% 21% 25% 012 21% 21% 055 23% 18% 035 16% 28% 068 19% 22% 021 185 30 17% 33%

Days from qualifying event to enrolment (median)

Old infarct in the territory of symptomatic artery

ICA=intracranial carotid artery. MCA=middle cerebral artery. PTAS=percutaneous transluminal angioplasty and stenting. TIA=transient ischaemic attack. *The product-limit estimate of the cumulative probability of a primary endpoint by 2 years after enrolment. p value for the interaction of treatment and the factor from a Cox proportional hazards model that included treatment, treatment by time interaction, factor, and treatment by factor interaction. As assessed by brain imaging after qualifying event.

Table 4: Comparison of treatment groups for the primary endpoint in pre-specied subgroups

2 years for the medical and PTAS groups are shown in table 4. We detected no interaction with treatment for any of the factors.

Discussion
The early benet of aggressive medical treatment compared with PTAS in high-risk patients with intracranial arterial stenosis persisted over a median duration of 324 months of follow-up in this trial. Although we cannot rule out the possibility that longer follow-up would have shown less benet from medical treatment alone, we think this possibility is unlikely because there was no suggestion that the ecacy gap between the two groups narrowed over timethe absolute risk reduction from medical treatment was 89% at 30 days and 90% at 3 years. These data imply that even if the 30-day rate of stroke and death had been as low as anticipated in the PTAS group (ie, similar to the
6

30-day rate of 58% in the medical group), PTAS would still not have provided any long-term benet over medical treatment alone because the anticipated benet of PTAS over medical treatment beyond the peri-procedural period did not occur. This nding was mainly because the primary endpoint rate in the medical group (141% at 2 years) was much lower than was projected (247% at 2 years). The much lower rate of stroke in the medical group in SAMMPRIS compared with patients in WASID who had similar entrance criteria,4,13,14 is probably explained by dierences in medical treatment in these trials. Patients in WASID were treated with usual risk factor management and either aspirin or warfarin, whereas patients in SAMMPRIS were treated with intensive risk factor management and combined aspirin and clopidogrel for 90 days followed by aspirin alone. In SAMMPRIS, SBP lower than 140 mm Hg and LDL-C lower than 181 mmol/L were attained in 34% (SBP) and 24% (LDL-C) of patients at enrolment and in 70% and 62% of patients at 1 year (gure 2), whereas in WASID these same risk factor levels were attained in 49% and 6% of patients at enrolment and in 50% and 12% of patients at 1 year (the LDL-C target in WASID was <259 mmol/L, which was achieved in 42% of patients at 1 year).15 At 30 days after enrolment, when the benets of intensive risk factor management would be expected to be limited, the stroke or death rate in the medical group in SAMMPRIS (58%) was almost half the rate of that in those patients in WASID with similar entrance criteria (107%).14 This nding suggests that dual antiplatelet therapy could have played an important part in lowering the early risk of stroke in the medical group of SAMMPRIS. Findings from two other trials have also shown that dual antiplatelet therapy can be more eective than aspirin alone in recently symptomatic patients with intracranial stenosis16 or at high risk of harbouring intracranial stenosis.17 Concerns have been raised that the medical management used in SAMMPRIS is not real-world practice and that the early success in controlling risk factors would not be sustained over long-term follow-up.18 However, all the essential elements of medical management in SAMMPRIS (dual antiplatelet therapy, intensive management of blood pressure and LDL-C, and the lifestyle programme) can be incorporated into clinical practice. Blood pressure and LDL-C were managed by study coordinators and neurologists who followed the studys risk factor management algorithms. These algorithms are available elsewhere7 and could easily be used in practice. Provision of a lifestyle management programme to patients with stroke is not yet routine care. However, an American Heart Association Presidential Advisory committee recommended that third-party payers should cover the costs of evidence-based lifestyle programmes for secondary prevention.19 Concern regarding the durability of risk factor control in

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

SAMMPRIS should be allayed by the data showing continued improvement in patients achieving targets for both primary (gure 2) and most secondary risk factors (appendix) over a median follow-up of 324 months. In our pre-specied subgroup analyses, we did not identify any subgroup that beneted from PTAS compared with medical treatment (table 4). The subgroup with a qualifying event while taking antithrombotic treatment constitutes the type of patients treated with Wingspan under the original Humanitarian Device Exemption in clinical practice. These patients had a worse outcome in the PTAS group compared with the medical group in SAMMPRIS (table 4). After publication of the early SAMMPRIS results,6 the FDA narrowed the Humanitarian Device Exemption criteria for use of the Wingspan stent to patients with 7099% stenosis who have had two or more strokes despite aggressive medical management, and whose last stroke occurred more than 7 days ago.20 We cannot provide specic data for this small subgroup of patients (data for medical management at the time of a stroke before the qualifying stroke were not collected), but the closest surrogate is shown in the last variable in table 4. Although this subgroup seems to be at very high risk of stroke on aggressive medical management, there was a similarly high rate of stroke in this subgroup treated with PTAS. Therefore, there are no data suggesting that PTAS with the Wingspan stent is more eective than medical treatment for any subgroup of patients with intracranial stenosis. However, if other endovascular techniques, such as angioplasty alone,21,22 can substantially lower the risk of the most frequent types of peri-procedural strokes that occurred in SAMMPRIS (ie, perforator stroke in the region of stent deployment, subarachnoid haemorrhage, and parenchymal brain haemorrhage related to reperfusion),23,24 further randomised trials assessing these techniques in high-risk patients might be needed. This study has some limitations. Because only one of the treatment groups underwent stenting, the trial could not be double masked. However, we tried to make the assessment of potential endpoints as unbiased as possible by making sure that a second site neurologist who was masked to treatment also assessed patients with potential mild stroke endpoints because these are the events that are most dicult to judge and subject to potential bias. Although we selected and credentialled experienced neurointerventionists to participate in this trial, it is possible that the high rate of peri-procedural stroke in SAMMPRIS in part reects the fact that intracranial stenting is still an uncommon procedure compared with coronary stenting, and that the complication rates of intracranial stenting will decrease with more experience. However, the challenge is to obtain this clinical experience without putting patients at excessive risk. Additionally, even if the peri-procedural risk could be reduced substantially, it is unlikely to be lowered to the level needed for intracranial stenting to be more

Panel: Research in context Systematic review This study (SAMMPRIS) is the rst randomised trial to compare aggressive medical management alone with aggressive medical management plus percutaneous transluminal angioplasty and stenting (PTAS) with the Wingspan stent system in patients with atherosclerotic intracranial arterial stenosis. Wingspan is the only stent approved by the FDA (under a humanitarian device exemption) for use in patients with atherosclerotic intracranial arterial stenosis. We know that SAMMPRIS is the rst randomised trial to assess the Wingspan stent from participating in an FDA panel meeting (March 23, 2012; Washington, DC, USA) on the Wingspan stent after the early results of SAMMPRIS had been published. We also searched Medline, the Cochrane Central Register of Controlled Trials, the Internet Stroke Trials Center Registry, and ClinicalTrials.gov from 2005 (when Wingspan was approved by the FDA) for publications in English and identied no other completed randomised trials using the Wingspan stent. We did our last search on Sept 15, 2013. Before SAMMPRIS began, three single-arm multicentre studies provided limited pilot data for the safety and potential ecacy of the Wingspan stent in patients with intracranial arteral stenosis and set the stage for the SAMMPRIS trial. Interpretation The nal results of SAMMPRIS show that aggressive medical treatment is superior to PTAS with the Wingspan stent in patients with recent transient ischaemic attack or stroke (within past 30 days) attributed to 7099% atherosclerotic intracranial arterial stenosis. PTAS was inferior because of the high risk of peri-procedural stroke and because there was no benet from PTAS beyond the peri-procedural period. Other key ndings in the study were as follows. First, high success rates at achieving target levels for vascular risk factors, particularly the primary risk factors (systolic blood pressure and low-density lipoprotein cholesterol) throughout the duration of the trial (median follow-up 324 months) using the studys risk factor protocols. This, in combination with the use of aspirin and clopidogrel for 90 days followed by aspirin alone, probably contributed to the much lower than expected risk of stroke in the medical group. Second, that the risk of stroke beyond 1 year in medically treated patients with severe intracranial arterial stenosis is low. And third, there was no suggestion that any subgroup of patients beneted from PTAS compared with medical management alone. Overall, these results lend support to the use of aggressive medical management rather than PTAS with the Wingspan system in high-risk patients with atherosclerotic intracranial arterial stenosis.

eective than aggressive medical management because of the inherent risk of the procedure. The nal results of SAMMPRIS show that aggressive medical treatment is superior to PTAS with the Wingspan system at both early and later phases of follow-up (panel). Although aggressive medical management seems to have substantially lowered the risk of stroke in these patients compared with historical controls managed with usual medical management, there are still subgroups of patients at high risk of stroke on aggressive medical treatment. Future research should focus on the development of new treatments to lower the risk of stroke in these high-risk patients. In this regard, it is interesting that the risk of stroke in the territory of a stenotic artery beyond 1 year of follow-up was very low in the medical group in SAMMPRIS, as was the case in WASID.4,13 A lower risk of stroke over time has also been shown in patients with symptomatic atherosclerotic stenosis (7099%) of an extracranial internal carotid artery.25 One possible
7

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

explanation for these ndings is that the common pathophysiological mechanisms of recurrent stroke in patients with atherosclerotic intracranial or extracranial stenosis (artery-to-artery embolism and distal hypoperfusion)2628 become attenuated over time. This might be related to stabilisation of vulnerable plaque29 and improvement in collateral blood ow,30 in which case new treatments that accelerate these reparative processes could become important components of the treatment of these high-risk patients.
Contributors CPD and MIC contributed equally. Both were involved in study conception and design, grant submission, protocol development, study oversight, data collection, data analysis and interpretation, and manuscript drafting and editing. MIC is the Principal Investigator (PI) for the SAMMPRIS trial. CPD was the NeuroInterventional PI initially and then became the NeuroInterventional co-PI after DF joined the study executive committee. MJL was involved in study conception and design, grant submission, protocol development, study oversight, data collection, data analysis and interpretation, and manuscript drafting and editing. MJL is the Biostatistics and Data Management PI. DF was involved in protocol development, data collection, study oversight, data analysis and interpretation, and manuscript drafting and editing. DF is the NeuroInterventional co-PI. TNT was involved in protocol development, data collection, study oversight, data analysis and interpretation, and manuscript drafting and editing. TNT is a member of the study executive committee in charge of the medical management regimen and risk factor reduction programme. LSJ was involved in study oversight as the NINDS representative on the study executive committee. In this capacity he oversaw data collection, study oversight, data analysis and interpretation, and manuscript editing. JM is a member of the study executive committee and serves as the project manager. In this capacity she oversaw data collection, study oversight, data analysis, and interpretation, and manuscript editing. AN participated in data collection and analysis. BFL was a former member of the study executive committee as the initial project manager. She participated in study design, grant submission, protocol development, data collection, study oversight, data analysis and interpretation, and manuscript editing. HLL and SLB are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis and interpretation. They edited this manuscript. MFW and BLH are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. LMH and EIL are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. AVA and MRH are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. DC and RPK are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. JMC and CGM are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. MDJ and GLP are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. JRL and OOZ are site neurology and neurointerventional PIs, respectively. They are members of the study steering committee and contributed to data collection, analysis, and interpretation. They edited this manuscript. ZR contributed to data collection, analysis, and interpretation in his role as a central neuroradiology reviewer of study brain imaging. He edited this manuscript. HJC contributed to data collection, analysis, and interpretation in his role as a central core laboratory for review of angiographic imaging. He edited this manuscript.

Conicts of interest CPD has associations with companies that manufacture medical devices for the treatment of cerebovascular disease in general, although none directly involved in this studythese are WL Gore and Associates (Scientic Advisory Board and Consultant), Microvention (Angiographic Core Laboratory for clinical trial), Penumbra (DSMB member for clinical trial), and Pulse Therapeutics (Chair, Scientic Advisory Board). MIC reports grants from NIH/NINDS and other support from AstraZeneca and Stryker Neurovascular (formerly Boston Scientic Neurovascular) related to this study. He also reports other grants from NIH/NINDS and personal fees from Gore Associates, Merck /Parexel, and Medtronic for participating as a stroke adjudicator or data safety monitoring board member on clinical trials unrelated to the submitted work. He also reports personal fees as an Expert Witness in medical legal cases related to stroke. MJL reports grants from National Institute of Neurological Disorders and Stroke during the conduct of the study. DF reports Siemens Microintervention Institutional payment for research and salary support; Covidien Ev3, Cordis, NFocus Medical, MicrusEndovascular consulting fees; Codman and Shurtle (REVIVE) royalties; Vascular Simulators LLC, TDC Technologies and CVSL ownership and stock interests. TNT reports grants from NIH/NINDS and other support from AstraZeneca and Stryker Neurovascular related to this study; a K23 grant from NIH/NINDS unrelated to this project; other from CardioNet (consultant); personal fees from Gore and Boehringer Ingelheim for participating as a stroke adjudicator in clinical trials unrelated to this work; and personal fees as an Expert Witness in medical legal cases unrelated to this research. AN reports grant from the National Institute for Neurological Disorders and Stroke (U01 NS058728). HLL reports one-time payments for investigator meetings and executive committee participation for this study. EIL reports grants from Boston Scientic Corporation, other from Codman and Shurtle, ev3/Covidien Vascular Therapies; Boston Scientic, other from Intratech Medical and Mynx Access Closure, other from Codman and Shurtle and TheraSyn Sensors, other from EV3 Covidien Vascular therapies & Blockade Medical LLC, other from Boston Scientic, other from Abbott Vascular and ev3 Covidien Vascular, other from Medical Legal Review outside the submitted work. CGM reports consultant and proctor for eV3, consultant for Microvention, and consultant for Codman. ZR reports NIH Federal Funding for this grant and other research support from Siemens and Bracco Diagnostics. All other authors declare that they have no conicts of interest. Acknowledgments This study was funded by a research grant (U01 NS058728) from the US Public Health Service National Institute of Neurological Disorders and Stroke (NINDS). Additionally, the following Clinical and Translational Science Awards, funded by the National Institutes of Health, provided local support for the assessment of patients in the trial: Medical University of South Carolina (UL1RR029882), University of Florida (UL1RR029889), University of Cincinnati (UL1RR029890), and University of California, San Francisco (UL1RR024131). Stryker Neurovascular (formerly Boston Scientic Neurovascular) provided study devices and supplemental funding for third party device distribution, site monitoring, and study auditing. This research was also supported by the InvestigatorSponsored Study Program of AstraZeneca, which donated rosuvastatin (Crestor) to study patients. INTERVENT provided the lifestyle modication programme to the study at a discounted rate. The Regulatory and Clinical Research Institute (RCRI; Minneapolis, MN, USA) provided assistance in designing the site monitoring processes and did the site monitoring visits. The VA Cooperative Studies Program Clinical Research Pharmacy Coordinating Center (Albuquerque, NM, USA) handled the procurement, labelling, distribution, and inventory management of the study devices and rosuvastatin. Walgreens pharmacies provided study drugs other than rosuvastatin to patients at a discounted price (paid for by the study). The PACE self-assessment forms for physical activity and smoking cessation were provided by the San Diego Center for Health Interventions, LLC. We thank the patients for participating in this study; Oscar Benavente, Carole White, Robert Hart, Pablo Pergola, and Ana Roldan of the Secondary Prevention of Small Subcortical Strokes trial (SPS3, NCT00059306) for assisting with the development and implementation of the SAMMPRIS blood pressure management protocol;

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3

Articles

George Howard and Thomas Brott for providing advice on study design and other issues based on their experience with the Carotid Revascularization Endarterectomy versus Stenting Trial (CREST, NCT00004732); and Rie Calcaterra who helped with the SAMMPRIS grant application and with the launching of the trial. References 1 Gorelick PB, Wong KS, Bae HJ, Pandey DK. Large artery intracranial occlusive disease: a large worldwide burden but a relatively neglected frontier. Stroke 2008; 39: 239699. 2 Arenillas JF. Intracranial atherosclerosis: current concepts. Stroke 2011; 42 (suppl): S2023. 3 Wong KS, Li H. Long-term mortality and recurrent stroke risk among Chinese stroke patients with predominant intracranial atherosclerosis. Stroke 2003; 34: 236166. 4 Kasner SE, Chimowitz MI, Lynn MJ, et al, and the Warfarin Aspirin Symptomatic Intracranial Disease Trial Investigators. Predictors of ischemic stroke in the territory of a symptomatic intracranial arterial stenosis. Circulation 2006; 113: 55563. 5 Chimowitz MI, Lynn MJ, Turan TN, et al, and the SAMMPRIS Investigators. Design of the stenting and aggressive medical management for preventing recurrent stroke in intracranial stenosis trial. J Stroke Cerebrovasc Dis 2011; 20: 35768. 6 Chimowitz MI, Lynn MJ, Derdeyn CP, et al, and the SAMMPRIS Trial Investigators. Stenting versus aggressive medical therapy for intracranial arterial stenosis. N Engl J Med 2011; 365: 9931003. 7 Turan TN, Lynn MJ, Nizam A, et al, and the SAMMPRIS Investigators. Rationale, design, and implementation of aggressive risk factor management in the Stenting and Aggressive Medical Management for Prevention of Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) trial. Circ Cardiovasc Qual Outcomes 2012; 5: e5160. 8 Bose A, Hartmann M, Henkes H, et al. A novel, self-expanding, nitinol stent in medically refractory intracranial atherosclerotic stenoses: the Wingspan study. Stroke 2007; 38: 153137. 9 US Food and Drug Administration. Summary of safety and probable benet. http://www.fda.gov/ohrms/dockets/ dockets/05m0308/05m-0308-aav0001-Summary-of-Safety-vol1.pdf (accessed Sept 10, 2013). 10 Samuels OB, Joseph GJ, Lynn MJ, Smith HA, Chimowitz MI. A standardized method for measuring intracranial arterial stenosis. AJNR Am J Neuroradiol 2000; 21: 64346. 11 Gordon NF, Salmon RD, Franklin BA, et al. Eectiveness of therapeutic lifestyle changes in patients with hypertension, hyperlipidemia, and/or hyperglycemia. Am J Cardiol 2004; 94: 155861. 12 Derdeyn CP, Fiorella D, Lynn MJ, et al, and the SAMMPRIS Trial Investigators. Impact of operator and site experience on outcomes after angioplasty and stenting in the SAMMPRIS trial. J Neurointerv Surg 2013; 5: 52833. 13 Chimowitz MI, Lynn MJ, Howlett-Smith H, et al, and the Warfarin-Aspirin Symptomatic Intracranial Disease Trial Investigators. Comparison of warfarin and aspirin for symptomatic intracranial arterial stenosis. N Engl J Med 2005; 352: 130516. 14 Zaidat OO, Klucznik R, Alexander MJ, et al, and the NIH Multi-center Wingspan Intracranial Stent Registry Study Group. The NIH registry on use of the Wingspan stent for symptomatic 70-99% intracranial arterial stenosis. Neurology 2008; 70: 151824. 15 Chaturvedi S, Turan TN, Lynn MJ, et al, and the WASID Study Group. Risk factor status and vascular events in patients with symptomatic intracranial stenosis. Neurology 2007; 69: 206368.

16

17

18

19

20

21 22

23

24

25

26

27

28

29

30

The CLAIR Study Investigators. The eectiveness of dual antiplatelet treatment in acute ischemic stroke patients with intracranial arterial stenosis: a subgroup analysis of CLAIR study. Int J Stroke 2012; published online Aug 7. DOI:10.1111/j.1747-4949.2012.00828.x. Wang Y, Wang Y, Zhao X, et al, and the CHANCE Investigators. Clopidogrel with aspirin in acute minor stroke or transient ischemic attack. N Engl J Med 2013; 369: 1119. Marks MP. Is there a future for endovascular treatment of intracranial atherosclerotic disease after Stenting and Aggressive Medical Management for Preventing Recurrent Stroke and Intracranial Stenosis (SAMMPRIS)? Stroke 2012; 43: 58084. Balady GJ, Ades PA, Bittner VA, et al, and the American Heart Association Science Advisory and Coordinating Committee. Referral, enrollment, and delivery of cardiac rehabilitation/ secondary prevention programs at clinical centers and beyond: a presidential advisory from the American Heart Association. Circulation 2011; 124: 295160. US Food and Drug Administration. FDA Safety communication: narrowed indications for use for the Stryker Wingspan Stent System. http://www.fda.gov/MedicalDevices/Safety/ AlertsandNotices/ucm314600.htm (accessed Sept 10, 2013). Marks MP, Wojak JC, Al-Ali F, et al. Angioplasty for symptomatic intracranial stenosis: clinical outcome. Stroke 2006; 37: 101620. Dumont TM, Kan P, Snyder KV, Hopkins LN, Siddiqui AH, Levy EI. Revisiting angioplasty without stenting for symptomatic intracranial atherosclerotic stenosis after the stenting and aggressive medical management for preventing recurrent stroke in intracranial stenosis (SAMMPRIS) study. Neurosurgery 2012; 71: 110310. Fiorella D, Derdeyn CP, Lynn MJ, et al, and the SAMMPRIS Trial Investigators. Detailed analysis of periprocedural strokes in patients undergoing intracranial stenting in Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS). Stroke 2012; 43: 268288. Derdeyn CP, Fiorella D, Lynn MJ, et al, and the Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis Trial Investigators. Mechanisms of stroke after intracranial angioplasty and stenting in the SAMMPRIS trial. Neurosurgery 2013; 72: 77795, discussion 795. Rothwell PM, Eliasziw M, Gutnikov SA, et al, and the Carotid Endarterectomy Trialists Collaboration. Analysis of pooled data from the randomised controlled trials of endarterectomy for symptomatic carotid stenosis. Lancet 2003; 361: 10716. Wong KS, Gao S, Chan YL, et al. Mechanisms of acute cerebral infarctions in patients with middle cerebral artery stenosis: a diusion-weighted imaging and microemboli monitoring study. Ann Neurol 2002; 52: 7481. Lee DK, Kim JS, Kwon SU, Yoo SH, Kang DW. Lesion patterns and stroke mechanism in atherosclerotic middle cerebral artery disease: early diusion-weighted imaging study. Stroke 2005; 36: 258388. Derdeyn CP. Mechanisms of ischemic stroke secondary to large artery atherosclerotic disease. Neuroimaging Clin N Am 2007; 17: 30311, viiviii. Fleg JL, Stone GW, Fayad ZA, et al. Detection of high-risk atherosclerotic plaque: report of the NHLBI Working Group on current status and future directions. JACC Cardiovasc Imaging 2012; 5: 94155. Liebeskind DS, Cotsonis GA, Saver JL, et al, and the Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) Investigators. Collaterals dramatically alter stroke risk in intracranial atherosclerosis. Ann Neurol 2011; 69: 96374.

www.thelancet.com Published online October 26, 2013 http://dx.doi.org/10.1016/S0140-6736(13)62038-3